Over the years, certain digital voltmeter (DVM) designs have improved input signal observation time but have also concurrently introduced sensitivity to the offset drift of the input amplifier feedback, thus limiting the performance of the product. For example, certain nanoVolt DVMs have used a measure zero cycle to remove the input amplifier's offset. In such DVMs, a measurement with the input applied to the input amplifier is typically made followed by a measurement with zero applied. The difference thus removes the input amplifier's offset but the zero cycle is time spent not measuring the input signal.
Other designs have used a reversing input amplifier that inverses its offset between each measurement. Thus, by averaging two adjacent measurements, the input amplifier's offset is cancelled and the input can be observed during both measurements thereby reducing the noise. However, the offset drift of the input amplifier's feedback is not cancelled and, therefore, limits the performance of the unit.
Accordingly, a need remains for a design that maintains high observation time while removing the detrimental sensitivity to the feedback offset drift.